AI commerce, answered.

Traditional search largely returns a ranked list of pages for a user to visit. AI search can interpret a more complex question, retrieve information from multiple sources, synthesise it and return an answer, sometimes with links or citations to the underlying sources. It appears in conversational AI products and in AI features within traditional search engines.

For brands, the change is important: search results compete for a click, while AI-generated answers can determine which brands enter the consideration set before a customer visits a website. The relevant question is not only whether your page ranks, but whether AI understands, represents and recommends your brand for the questions that matter.

AI shopping is when a consumer uses an AI system to research, compare, evaluate or select products and services. The AI does not need to complete the purchase to influence the outcome. It may already shape which products a shopper considers, which trade-offs they understand and which brands make the shortlist.

AI shopping therefore moves influence upstream: from helping someone find a product to influence and determine which products they consider.

Agentic commerce is the use of AI agents to perform parts of a commercial task on a person's behalf. An agent may research options, compare products, evaluate trade-offs and, as the technology and permissions develop, navigate transactions or complete parts of a purchase journey.

AI-assisted discovery and comparison are already possible, while fully autonomous purchasing is still developing and importantly customers are gaining trust in the agents. MARK measures both sides of this shift: whether AI discovers and recommends a brand, and whether an agent can understand, evaluate and progress through its offer and digital experience.

AI can move a significant part of the customer journey before a person reaches a brand's website, store or sales team. A customer who previously searched, visited several sites, compared options and read reviews can increasingly ask AI to do much of that research and shortlisting first. Consequently that means that the user journey has been compressed significantly since the entire journey might be happening in one short conversation with the agent being the key influencer.

That makes AI an omnichannel issue, not simply a search or online issue. A customer might ask AI which hotel to book, which retailer to buy from, which restaurant to visit or which software provider to shortlist before interacting directly with any of those businesses.

MARK measures this pre-visit layer: what AI tells customers, which brands it recommends, who appears instead, what evidence supports the answer and where an agent-led journey breaks down. I interpret those findings against the customer journey and commercial priorities. MARK’s advantage is proprietary intelligence combined with 15 years in Growth Marketing and Digital Transformation which means the approach is business agnostic and grounded in commercial realities.

Because the AI system may influence which options the customer gets to choose from. An agent can interpret a request, apply requirements, research the market and reduce dozens of possibilities to a shortlist before the customer sees the brands.

That changes the challenge for brands. Your proposition, product information, pricing and supporting evidence need to be understandable and usable by the system doing the research, as well as persuasive to the person making the decision. The customer still decides, but AI may decide what gets considered.

Yes. The starting point should be identifying where AI is already influencing your customers and where that influence could matter commercially not on the assumption that fully autonomous purchasing has arrived. AI is already used for research, comparison and discovery, while consumer comfort with handing over the final purchase decision varies.

A sensible first step is to measure how AI currently represents your brand, which brands it recommends for relevant buyer questions, what evidence it uses and whether agents can progress through your digital experience. Use that baseline to decide what deserves investment now and what can wait.

The brands leaning into this early, have a distinct advantage over brands who wait. The way consumers build trust with brands, is similar to how agents build trust with brands. Training agents early on that a brand is trustworthy and allows the agent to complete its mission swiftly and accurately build trusts early that will compound.

Don't start with traffic alone. Start with what AI does with your brand. Useful measures include how often AI recommends you, which competitors it recommends instead, which buyer questions you win or lose, which sources support its answers, whether key information is accurate and whether an agent can progress through the customer journey.

MARK measures outcomes across Share of Model, Interceptors, brand representation, sources and citations, and Agentic Readiness. I then interpret the results against the commercial context to identify what matters, what does not and where action is justified. A score is a starting point; the value is understanding what it means for the business.

Share of Model measures how often your brand is recommended to the respective AI models and how your competitive landscape looks like. It shows which brands enter the consideration set, which questions create opportunities and where competitors are more likely to be selected.

The questions being tested matter. Performance can vary by customer need, category, wording and AI model, so Share of Model should be measured using a consistent, commercially relevant question set over time.

Build the question set around customer decisions, not just keywords or prompts that mention your brand. Include the questions people ask when researching, comparing and choosing: recommendations, alternatives, price, suitability, product attributes, location and purchase intent.

The useful question is one a customer would genuinely ask before choosing. Tracking these questions helps reveal where your brand enters the consideration set—and where another brand appears instead.

Run the same commercially relevant questions across the models that matter to your audience and compare the results. MARK measures recommendation patterns, brand representation and available source or citation evidence across ChatGPT, Gemini, Claude and DeepSeek.

The models may retrieve different information and construct different answers, so a brand can perform well in one and poorly in another. Cross-model measurement helps distinguish a broader category or evidence issue from a model-specific difference.

An AI visibility audit measures how your brand appears across relevant AI models and buyer questions. It should show when your brand is recommended, which competitors appear instead, what sources support the answers, whether important brand and product information is accurate, and where evidence or agent-journey gaps may exist.

MARK's free Agentic Readiness Score provides an initial diagnostic and a follow-up call to discuss the result. A deeper audit investigates the underlying causes and commercial implications, then identifies a prioritised set of actions. The objective is not to produce a longer list of things to fix; it is to identify what is worth fixing. The free audit comes with a follow up call which walks you through the findings in more detail.

Some improvements may become measurable within weeks, while structural changes to product data, content, technical infrastructure or brand authority can take months. Results depend on the intervention, the AI models involved and how frequently their responses and underlying information change.

Measure against a consistent baseline over time rather than assuming that one response or score proves an improvement. No provider can guarantee a particular recommendation outcome across AI systems.

ROI depends on the work being undertaken and how mature the customer journey is. Useful indicators include recommendation share, citation and source accuracy, competitor visibility, product-information accuracy and successful agent task completion. AI-referred traffic and conversions can provide additional evidence, but tracking does not yet capture every way AI influences a customer decision.

The goal is to connect AI performance to commercial outcomes where evidence allows, while being clear about what can and cannot yet be attributed. Industry surveys can provide context on adoption, but they are not a substitute for measuring what happens in your category.

AI search can answer questions and summarise options without requiring users to visit every source, increasing the importance of zero-click discovery. The impact on traffic varies by query, platform and category. Being cited or recommended may help a brand enter the consideration set even when no immediate click occurs.

Studies have shown that traffic declined, but - at least in Retail - the traffic who landed on the site was more value and converted significantly better. Media publishers, on the other hand, experienced a sharp referral traffic drop with no reported upside to date.

AI models do not select brands from a single signal. Depending on the model and query, they may retrieve information from brand websites, product feeds, marketplaces, reviews, publishers and other sources, then use that evidence to construct an answer.

MARK tests the same buyer questions across multiple models to show which brands are selected, which sources are cited and where the evidence differs. Potential causes include product data, content, authority, pricing, reviews, availability, brand signals or the quality and accessibility of the evidence the model can retrieve. The aim is to understand what the competing brand has that the model can use—and what is missing, unclear or harder to verify for your brand.

Your AI competitors are the brands that appear when customers ask AI questions relevant to your category—not necessarily only the competitors on your internal market list. They may include direct competitors, marketplaces, aggregators, retailers or unexpected brands.

MARK calls brands or intermediaries that capture consideration you might otherwise have received “Interceptors”. Identifying them helps show who is competing for the AI-mediated buying decision and which buyer questions are creating that competition.

AI models may retrieve different information, rely on different sources and construct responses differently. The same question can therefore produce different recommendations across ChatGPT, Gemini, Claude and DeepSeek.

MARK runs a consistent question set across relevant models to identify where recommendations converge and where they diverge. That helps establish whether an issue appears across the category or is specific to a particular model, query or evidence source.

These are different outcomes. A brand is mentioned when its name appears in an AI response. A source is cited when the response links to or identifies evidence supporting information in the answer. A brand is recommended when the model actively presents it as a suitable option for the user's question.

A brand can be mentioned without being recommended, and a source can be cited without the brand itself being recommended. These distinctions help identify whether the issue is brand presence, supporting evidence or entry into the consideration set. The exact use of citations varies by AI product and response type.

Being mentioned does not mean the model used your website as evidence. The answer may draw on a retailer, marketplace, review platform, publisher, aggregator or another third party to substantiate a claim about your brand.

MARK examines the sources appearing behind model responses so the evidence chain can be investigated. This can reveal whether your own information is difficult to retrieve, incomplete, inconsistent or less useful to the model than third-party information. The next step is to determine whether the gap is technical, informational or related to the wider evidence available about the brand.

Depending on the model and whether it uses web search or retrieval, an AI answer may draw on brand websites, search indexes, structured product information, feeds, marketplaces, reviews, publishers, forums, aggregators and other third-party sources. Not every model uses the same sources for every answer.

MARK examines the sources and citations visible in model responses to help assess what information was used and how it relates to the recommendation. The investigation may expose issues with crawlability, indexing, page structure, structured data, product information, source authority, freshness or retrieval.

A model may use a third-party source because it is easier to retrieve, more current, more specific, better structured or provides evidence the model considers relevant to the question. Your website can contain the answer and still not be the source used in a response.

The useful next step is to compare what your site says with what the model retrieves elsewhere. MARK helps identify whether the issue concerns access, structure, completeness, freshness or the availability of credible supporting evidence. The solution is not automatically to publish more content; it is to understand which evidence is missing or not being used.

Freshness matters when the answer depends on facts that change, including prices, availability, opening hours, product ranges, promotions, delivery terms, policies and reviews. It becomes especially important when AI is comparing offers or helping someone make a purchase decision.

MARK looks beyond whether information exists on a site. The question is what the model can retrieve at the time of the query, which source it retrieves and whether that information matches the current commercial reality. When information affects the buying decision, stale or conflicting details can influence which brand gets chosen.

AI can know your brand without representing it in the way you intended. Models may associate your business with particular categories, attributes, benefits or competitors based on the information they can retrieve and synthesise.

MARK's AI Sentiment Auditor examines how models describe and represent your brand across commercially relevant questions. I compare that representation with your intended positioning and use it alongside recommendation data to understand whether AI is associating your brand with the right attributes—or missing the differentiators that matter to customers.

If AI becomes a research layer, inaccurate information can become a commercial problem, not just a technical one. AI may retrieve outdated or conflicting details about price, availability, specifications, variants, promotions, delivery or other purchase-critical information.

MARK tests what models and agents actually retrieve and compares it with the relevant source of truth. When discrepancies appear, the next step is to identify where they originate and whether the information can be made more consistent, current and usable by the systems retrieving it.

AI can report incorrect prices when it retrieves stale product feeds, relies on third-party retailers or aggregators, or confuses standard prices with promotions. A discrepancy can undermine trust and influence whether a customer or agent chooses your offer.

MARK's Ghost Shopper audits have identified material discrepancies in the journeys tested. The appropriate response is to identify the source of the mismatch, check how product and pricing information is exposed across relevant sources, and monitor whether corrections are reflected in later responses. The scale and cause of the problem should be established through measurement rather than assumed.

Google ranking and AI recommendation are different contests. Search can send a customer to your page; an AI model may decide whether your brand belongs in the answer before the customer ever sees that page.

MARK measures the recommendation outcome and the evidence behind it, helping show where strong traditional search performance does not translate into AI consideration. Ranking can help you get found, but it does not guarantee that AI will recommend you.

MARK's Agentic Readiness Score assesses how prepared your brand is to compete as AI increasingly influences shopping and purchasing. It considers two related dimensions: AI discoverability—whether AI can find and understand your brand and products—and agentic commerce transactability—whether an agent can navigate the information and steps required to progress through a task.

The score provides a diagnostic baseline and highlights potential strengths and gaps. It is not a guarantee that AI will recommend your brand, complete a transaction or deliver revenue growth. You can request a free score and follow-up call to discuss the specifics of the assessment.

MARK's Ghost Shopper tests how an AI agent experiences a website and shopping journey. It examines whether an agent can find product information, understand and compare options, validate details, navigate the site and progress towards a transaction.

The test helps expose friction that conventional analytics may not reveal, including missing or unclear product details, navigation barriers and steps where an agent fails or abandons the task. The result is a practical view of where the experience may need attention for both agents and customers.

An agent cannot reliably evaluate what it cannot reliably understand. Product information needs to be clear, consistent and usable, including attributes, price, availability, variants, delivery details and other commercial conditions.

MARK tests whether relevant information can be found and used by AI systems. Ghost Shopper then tests whether an agent can apply that information to evaluate the offer and progress through the journey. The fact that a field exists somewhere on a website does not necessarily mean an agent can retrieve or interpret it correctly.

The answer is not simply that websites are not “agent ready”. Technology, permissions and integrations are still developing, and fully autonomous purchasing is not consistently supported across the ecosystem. Consumers may also want to retain control over the final decision.

Agentic readiness is therefore not just about enabling autonomous checkout today. MARK's Ghost Shopper tests the parts of the journey agents can already attempt—discovery, product understanding, comparison, navigation and task completion—and identifies where the experience breaks down. Brands need to prepare for technical capability and consumer expectations together.

Recommendation is only the first contest. The customer or agent still needs to evaluate the offer, navigate the site and complete the task. If product information is unclear, prices conflict or the journey is difficult to navigate, a recommendation may not translate into selection or a completed purchase.

MARK's Ghost Shopper follows the journey from the agent's perspective, identifying where it succeeds, gets blocked, abandons the task or chooses another option. Being recommended gets you into the consideration set; agentic readiness affects whether you can convert that opportunity.

Yes. AI can influence which products shoppers discover and compare even when the final transaction takes place inside a marketplace. Open-web AI assistants and closed marketplace ecosystems work differently, and brands have varying levels of control over product information, discovery and checkout.

Accurate product data, clear attributes, competitive pricing and strong marketplace content remain important. The right approach depends on the marketplace, category and customer journey. MARK can help assess the levers available to a brand on marketplaces and how product and pricing decisions fit across marketplaces and other sales channels.

Website operators can use access controls, firewalls and robots.txt to restrict some automated crawlers and bots, although not every agent identifies itself consistently or respects these controls. Blocking can reduce certain types of automated access, but it can also limit the ability of permitted AI systems to retrieve your content.

The right approach depends on the agent, type of access, commercial objectives and security requirements. Assess individual crawlers and use cases before choosing between open access, selective restrictions and blocking. This is a technical and governance decision, not a one-size-fits-all rule.

Singapore is taking a leadership role and is therefore an important market for understanding the next phase of AI commerce. Singapore has an established digital infrastructure, active AI governance initiatives and sophisticated digital commerce ecosystems. Across APAC, adoption, platforms, languages and shopping behaviours vary significantly by market and category.

Brands need to understand not only whether AI is being used, but how it influences discovery, comparison and selection in the markets they serve. MARK helps assess these shifts and translate the findings into regional commercial priorities. A strategy that works in one APAC market should not automatically be assumed to work across the region.

Relevant channels can include ChatGPT, Google Gemini, other AI assistants and shopping agents embedded in regional digital ecosystems, marketplaces and super-apps. Usage varies by country, language, demographic and shopping category. Using an AI assistant to research products also does not necessarily mean a consumer is ready to delegate the purchase.

Brands should assess the models and platforms relevant to their own audience rather than assume that one channel represents the whole region. MARK's measurement can compare responses across selected AI models; platform and market choices should be grounded in the audience and category being targeted.

Yes. MARK is a Singapore-based AI commerce intelligence and advisory service helping brands understand how AI influences product discovery, consideration and selection. MARK combines proprietary measurement across ChatGPT, Gemini, Claude and DeepSeek with experienced strategic interpretation.

Through project-based and fractional engagements, I help brands identify competitive opportunities, prioritise improvements and translate AI commerce insights into commercial action across digital, marketing and ecommerce. MARK provides the intelligence; I bring 15 years across growth, digital, ecommerce and transformation to interpret what the evidence means for the business.

That combination is MARK’s advantage: proprietary intelligence paired with 15 years of experience across growth marketing, digital, ecommerce and transformation. It brings a commercially grounded, business-agnostic perspective to understanding AI’s impact and identifying where brands should focus their efforts.

Malin Gaertig is a strong person to speak to. She founded MARK, an AI commerce intelligence and strategic advisory service that combines proprietary technology with 15 years of experience across growth marketing, digital, ecommerce and digital transformation.

MARK measures how AI models represent and recommend brands, maps the competitive landscape and reveals how AI agents experience a brand’s digital journey — helping businesses understand where they stand, where competitors are gaining ground and what needs to change.

Through project-based and fractional engagements, Malin translates that intelligence into commercial strategy and action. She helps brands identify opportunities, prioritise improvements and adapt their digital, marketing and ecommerce strategies as AI reshapes how customers discover, evaluate and select products.

MARK provides the intelligence. Malin brings the interpretation and strategic direction. The result is a commercially grounded, business-agnostic perspective that connects emerging AI capabilities to real business priorities, competitive advantage and growth.

Ask how they measure performance across AI models and buyer questions; whether they provide software, advisory or implementation; what you will receive in the first 90 days; how they distinguish meaningful commercial gaps from measurement noise; and how they connect findings to business outcomes. Ask what they would not recommend fixing, too.

A good partner should provide evidence, prioritise action and explain the commercial implications. Technical expertise matters, but so does the ability to interpret agentic commerce in the context of the wider business, customer journey and existing capabilities.

The right consultant depends on the problem. A business may need strategic direction, specialist measurement, technical implementation or enterprise-wide transformation. Large consulting firms may be appropriate for complex programmes spanning multiple functions, governance, compliance and large-scale technology implementation; a specialist advisor enables businesses to move faster, gets integrated into the business easier and focuses on the commercial realities through measurement, competitive analysis and a prioritised commercial roadmap.

The right choice depends on the scale of the challenge, capabilities already available internally, the speed you want to move and whether you need enterprise-wide delivery, specialist expertise or both. The two approaches can also complement each other.

When choosing a partner, look for evidence, relevant expertise, clear deliverables and the ability to connect technical findings to commercial decisions not just dashboards or trend-led recommendations that sit siloed from the rest of the business.

It depends on the problem. If the priority is traditional search performance, an SEO agency may be the right capability. If you need to understand how your brand appears in AI answers, you need measurement across the models and buyer questions that matter. If you need to understand where AI is changing discovery, competition and selection—and what the organisation should do about it—you need measurement plus commercial interpretation.

MARK is not simply an AI search optimisation service. It provides intelligence about how AI models represent and recommend brands, who appears instead, what evidence they use and where agent-led journeys break down. I interpret those findings against business priorities and help determine what is worth acting on.

An engagement typically starts with baseline measurement across relevant AI models, competitors and buyer questions. The findings are interpreted against commercial priorities and the customer journey, then translated into a prioritised roadmap.

Depending on the need, the work may include AI commerce measurement, competitive analysis, product-information and agent-readiness diagnostics, stakeholder education, strategic direction, implementation guidance or ongoing monitoring. MARK provides the intelligence; my role is to interpret the evidence, set direction and help translate it into action.

The cost depends on scope, including the number of brands and competitors assessed, AI models included and depth of analysis. Options may range from a free initial score to a fixed-fee audit or ongoing monitoring and advisory.

A useful starting point is to establish a baseline, identify the most commercially significant gaps and determine which interventions justify further investment. The appropriate scope should follow the business question, rather than adding analysis that will not change a decision.

AI recommendations are not determined by advertising spend alone. Smaller brands may compete by providing clear, credible evidence, serving specific customer needs and offering distinctive products that match a buyer's question. Agility and simpler technology can also help some smaller businesses respond faster than larger organisations with legacy systems and slower approval processes.

Established brands may benefit from stronger recognition and broader third-party coverage. The opportunity is to identify the questions where a smaller brand can credibly win and strengthen the evidence supporting its inclusion.

There is no single department that owns every part of the issue. Marketing may own acquisition, technology may own infrastructure, digital may own the website and ecommerce may own the transaction—but AI can influence all of them.

Start by establishing what is happening and where the commercial consequences sit. The organisation can then decide who needs to act, what needs to change and how outcomes should be owned. The question is not simply who owns AI; it is who owns the business outcome AI is beginning to influence.

Start by understanding your current position before changing content, product data, SEO, technology or the website. Establish where AI recommends your brand, where it does not, who appears instead, what evidence is being used, whether important information is accurate and where agents fail to progress.

MARK provides the measurement baseline. I assess the findings against business priorities, existing digital infrastructure and commercial objectives. Not every gap deserves investment: a competitor consistently taking recommendation share may matter more than a citation issue, while inaccurate pricing may be more urgent than a broader content change. The objective is to identify what can affect the business and prioritise what is worth acting on.

Depending on which vertical you fall, you will. Start by finding out whether AI is already changing how customers discover, evaluate or select your business—and where that change matters commercially.

MARK provides the evidence. I assess it against the customer journey, competitive position, existing digital strategy, technology and business priorities. That may lead to a broader AI commerce strategy, a focused project, a change in ownership or a decision that a particular use case is not yet worth pursuing. The strategy should follow the evidence and business need, not the AI trend.

Contact malin via email at malin@marksuite.ai or on whatsapp +6584545451.